102_satwind

bufr

El viento derivado de satélites se codifica como tipo 240 - 260 dependiendo del instrumento. Sin embargo ni GFS ni el resto de los sistemas operativos usan las observaciones presentes en el prepbufr y en cambio leen el bufr específico gdas.tHHz.satwnd.tm00.bufr_D.

En la Tabla 18 se detalla el uso de las distintas observaciones según el instrumento.

Satwnd Header Definitions:

• DATACAT = BUFR Data Category
• SAID = Satellite Identifier
• OCID = Identification of originating/generating center (GCLONG in BUFR)
• CLAS = Satellite Classification (SCLF in BUFR)
• RPID = Report Identifier
• SIDP = Satellite Used in Instrument Processing
• SWCM = Satellite Derived Wind Computation Method
• YEAR = YEAR
• MNTH = MONTH
• DAY = DAY
• HOUR = HOUR
• MIN = MIN
• LAT = Latitude (Deg)
• LON = Longitude (Deg)
• PRSL = Pressure (Pa)
• TMP = Temperature/Dry Bulb Temperature (K)
• WDIR = Wind Direction (Degree True)
• WSPD = Wind Speed (m/s)

Mejor y con más detalle Tabla b

GSI

read_obs.f90

• Reconoce el bufr de este tipo de observaciones si el archivo se llama satwndbufr y llama a la rutina read_satwnd

read_satwnd.f90

For the satellite ID type:

• 240: GOES short wave winds,
• 241: India,
• 242: JMA Visible,
• 243: EUMETSAT visible,
• 244: AVHRR winds
• 245: GOES IR.
• 246: GOES WV cloud top,
• 247: GOES WV deep layer
• 250: JMA WV deep layer.
• 251: GOES visible,
• 252: JMA IR winds
• 253: EUMETSAT IR winds,
• 254: EUMETSAT WV deep layer winds
• 257,258,259: MODIS IR,WV cloud top, WV deep layer winds
• 260: VIIR IR winds

For satellite subtype:

• 50-80 from EUMETSAT geostationary satellites(METEOSAT)
• 100-199 from JMA geostationary satellites(MTSAT)
• 250-299 from NESDIS geostationary satellites(GOES)
• 700-799 from NASA Terra and Aqua satellites
• <10, 200-223 from NOAA-15, 16, 17, 18, polar orbit and EUMESAT MetOp satellites

The quality mark:the values range from 0 to 15.

• 0-7 used: 0 is best, when the value greater than 3, the observation error needed to be enflated.
• 8-15 monitored

Notas:

• Rechaza observaciones por encima de 125mb.
• Rechaza observaciones con quality marker 12 o 14.
• Algunas restricciones extras dependiendo del tipo de observación, por ejemplo:
  • Velocidad mínima de 10m/s para CAWV (type = 247), qm = 15
  • qm=15 !reject data with low QI

diagfile

• cdata_all(1,iout)=woe ! wind error
• cdata_all(2,iout)=dlon ! grid relative longitude
• cdata_all(3,iout)=dlat ! grid relative latitude
• cdata_all(4,iout)=dlnpob ! ln(pressure in cb)
• cdata_all(5,iout)=ee ! quality information
• cdata_all(6,iout)=uob ! u obs
• cdata_all(7,iout)=vob ! v obs
• cdata_all(8,iout)=rstation_id ! station id
• cdata_all(9,iout)=t4dv ! time
• cdata_all(10,iout)=nc ! index of type in convinfo file
• cdata_all(11,iout)=qifn +1000.0_r_kind*qify ! quality mark infor
  
• cdata_all(12,iout)=qm ! quality mark
• cdata_all(13,iout)=obserr ! original obs error
• cdata_all(14,iout)=usage ! usage parameter
• cdata_all(15,iout)=idomsfc ! dominate surface type
• cdata_all(16,iout)=tsavg ! skin temperature
• cdata_all(17,iout)=ff10 ! 10 meter wind factor
• cdata_all(18,iout)=sfcr ! surface roughness
• cdata_all(19,iout)=dlon_earth_deg ! earth relative longitude (degrees)
• cdata_all(20,iout)=dlat_earth_deg ! earth relative latitude (degrees)
• cdata_all(21,iout)=zz ! terrain height at ob location
• cdata_all(22,iout)=r_prvstg(1,1) ! provider name
• cdata_all(23,iout)=r_sprvstg(1,1) ! subprovider name
• cdata_all(25,iout)=var_jb ! non linear qc parameter

Comparación analisis con y sin satwnd

Prueba de asimilación usando 3DVar para hacer un par de experimentos rápidos.

Fecha: 2018112100

exp[bottom_top %in% seq(1, 60, 10)] %>% 
  ggplot(aes(west_east, south_north)) +
  geom_point(aes(color = u)) +
  scale_color_divergent() +
  facet_grid(exp~bottom_top)

diff[bottom_top %in% seq(20, 58, 2)] %>% 
  ggplot(aes(west_east, south_north)) +
  geom_point(aes(color = u.diff)) +
  scale_color_divergent() +
  facet_wrap(~bottom_top)

Diagfiles

diag <- fread("/home/paola.corrales/comGSIv3.7_EnKFv1.3/examples/test/gsi_satwnd/results_conv_ges.2018112100") %>% 
  .[, c("V2", "V4") := NULL]

colnames(diag) <- c("var", "stationID", "type", "dhr", "lat", "lon", "pressure", "usage.flag", "obs", "obs.guess", "obs2", "obs.guess2", "rerr")

diag <- diag[type %between% c(240, 260)] %>% 
.[, sat_type := case_when(
  type == 240 ~ "GOES SW winds",
  type == 241 ~ "India",
  type == 242 ~ "JMA Visible",
  type == 243 ~ "EUMETSAT visible",
  type == 244 ~ "AVHRR winds",
  type == 245 ~ "GOES IR", 
  type == 246 ~ "GOES WV cloud top", 
  type == 247 ~ "GOES WV deep layer",
  type == 250 ~ "JMA WV deep layer", 
  type == 251 ~ "GOES visible", 
  type == 252 ~ "JMA IR winds",
  type == 253 ~ "EUMETSAT IR winds", 
  type == 254 ~ "EUMETSAT WV deep layer winds",
  type == 257 ~ "MODIS IR",
  type == 258 ~ "MODIS WV cloud top",
  type == 259 ~ "MODIS WV deep layer winds",
  type == 260 ~ "VIIR IR winds")] %>% 
  melt(measure.vars = c("obs", "obs2", "obs.guess", "obs.guess2")) %>% 
                       .[, var := if_else(str_detect(variable, "2"), "v", "u")] %>% 
                       .[, variable := str_remove(variable, "2")] %>% 
  pivot_wider(names_from = variable, values_from = value) %>% 
                       setDT

diag[var == "u"] %>% 
  ggplot(aes(ConvertLongitude(lon), lat)) +
  geom_point(aes(color = sat_type)) +
  geom_mapa() +
  coord_sf(xlim = c(-75, -52), ylim = c(-42,-20)) +
  facet_wrap(~usage.flag, labeller = labeller(usage.flag = c("-1" = "rejected",
                                                             "1" = "asimilated"))) +
  labs(color = "Source",
       x = "lon") +
  theme_minimal()

N_obs <- diag[var == "u", .N, by = .(usage.flag, var)] %>% 
  .[, label := paste("N = ", N)] %>% 
  .[, ":="(lon = 302.0,
          pressure = 920)]

diag[var == "u"] %>% 
  ggplot(aes(ConvertLongitude(lon), pressure)) +
  geom_point(aes(color = sat_type)) +
  scale_y_level(name = "pressure level", breaks = c(1000, 900, 700, 500, 400, 300, 200, 100)) +
  scale_x_longitude(breaks = seq(-75, -55, 5)) +
  geom_label(data = N_obs, aes(x = ConvertLongitude(lon), y = pressure, label = label))  +
  facet_grid(var~usage.flag, labeller = labeller(usage.flag = c("-1" = "rejected",
                                                             "1" = "asimilated"))) +
  labs(color = "Source",
       x = "lon",
       y = "pressure level") +
  theme_minimal()

diag %>% 
  ggplot((aes(ConvertLongitude(lon), lat))) +
  geom_point(aes(color = obs)) +
  scale_color_divergent() +
  geom_mapa() +
  coord_sf(xlim = c(-75, -52), ylim = c(-42,-20)) +
  facet_grid(var~usage.flag, labeller = labeller(usage.flag = c("-1" = "rejected",
                                                             "1" = "asimilated"))) +
  theme_minimal()

diag[var == "u"] %>% 
  ggplot(aes(ConvertLongitude(lon), pressure)) +
  geom_point(aes(color = obs)) +
  scale_color_divergent() +
  scale_y_level(name = "pressure level", breaks = c(1000, 900, 700, 500, 400, 300, 200, 100)) +
  scale_x_longitude(breaks = seq(-75, -55, 5)) +
  geom_label(data = N_obs, aes(x = ConvertLongitude(lon), y = pressure, label = label))  +
  facet_grid(var~usage.flag, labeller = labeller(usage.flag = c("-1" = "rejected",
                                                             "1" = "asimilated"))) +
  labs(color = "Source",
       x = "lon",
       y = "pressure level") +
  theme_minimal()